Re: [MiNT] Some mintlib patches

[Alan Hourihane <alanh@fairlite.demon.co.uk>]

On Thu, 2008-02-21 at 17:05 +0100, MiKRO wrote:
>   
>         I'm confused why Miro is still changing crt0.S ?
> We're talking about two different things now:

Right, which is what I thought and Keith confused me.

> 1. normal stack, which is reported by 'stack' command -- this one
> needs to be changed because the standard MINKEEP (64K I believe) is
> too low -- and as Keith says, you can't change cc1's stack size after
> crash since gcc checks crc. This wasn't included in my patch but I use
> MINKEEP = 1 MB for avoiding such crashes and for no need to catch when
> gcc fails -- just 'make' and wait until everything is done. That's the
> reason why I like changing MINKEEP with as little influence as
> possible, Alan ;)

I'm using MINKEEP == 64K at the moment and suffer no gcc crashes at all.

> 2. preliminary :) stack -- by default, this stack is as big as
> possible (after bss to hi tpa) but due to problems I write you in the
> previous mail, it doesn't work for some gcc binaries either, that
> means it crashes again. So I changed this value to some explicit one
> to be sure I know what's going on (and what is the cause, at least
> rough idea) -- and as you see, I discovered 1 KB is really too little
> for this preliminary stack. So I changed it to 64K (no relation to
> MINKEEP, though) and things are working again.

Right.

I'm off now to compile gcc 4.2.3 with MINKEEP set at 64K and the
original crt0.S and will report back what I find.

Alan.